The planning and consent stages of any wave farm development involve both resource and environmental impact assessments. Nearshore wave models such as SWAN can potentially be used for both these processes. For a resource assessment, wave states recorded at locations other than the proposed site, or at a substantial distance from the location of interest within a large site, can be propagated to the required location using the model, thus potentially eliminating the costs involved in additional wave recording devices. An estimate of potential impacts on the nearshore wave climate can be achieved by propagating wave states from an offshore boundary, through the wave farm site, to the nearshore region. By performing a model study with and without some form of wave farm representation, nearshore changes can be estimated.
It is common practice, when utilising a model such as SWAN for a specific region, to calibrate model coefficients for processes such as whitecapping, bottom friction and non-linear interactions in order to minimise differences between model output and recorded wave data. However, calibration can be a lengthy process and requires concurrent offshore and nearshore recorded datasets. Most offshore wave farms are likely to be sited in water depths greater than 40m, and for many wave states this would be considered a “deep-water” site. The necessity for a calibrated nearshore model then becomes questionable, particularly when considering parameters relevant to shallow water effects.
This paper investigates the potential gains in model output accuracy that could be achieved with calibration and thus the need for such a process, for both a resource assessment and an assessment of the impact of a wave farm on the nearshore wave climate. A particular focus of the investigation is the Wave Hub wave farm site in the southwest of the UK. A SWAN model calibration for this region is described, and both resource and nearshore wave climate impact assessments are performed in order to compare model output when using default coefficient values with that for calibrated values. The results indicate how the need for calibration for the purposes of resource assessment becomes more critical as water depth decreases. They also illustrate that a comparative study, where differences are calculated between the output of near-identical models, is almost unaffected by calibration and that this process could be removed from future studies.